Dead Heat

For those who are waiting for a review of the recent work on MSU temperature estimates, it’s coming. But it turns out to be a rather complex subject, with many papers to review and a lot of data to examine, so it’ll take a bit longer. In the meantime, the NASA GISS year-end estimates of monthly average temperature have been released, so let’s take a look. GISS provides monthly average temperature data from 1880 to the present for four areas: the land+sea (the globe), land, northern hemisphere land, and southern hemisphere land. The most oft-referenced data, which represent land and ocean for the entire planet, are the “globe” (land+sea) numbers.

I stated earlier that globally, 2007 was likely to end up being the 2nd-warmest year in the GISS record, surpassing 1998. It turns out that December 2007 was cooler than expected, so those two years ended in an exact tie; they’re in a “dead heat” for 2nd place. The hottest year in the GISS record is 2005; the five hottest years, in order, are:

Place	Year	Average
1st	2005	0.62
2nd-3rd	2007 (tie)	0.57
2nd-3rd	1998 (tie)	0.57
4th	2002	0.56
5th	2003	0.55

Here’s the time series plot of annual average temperature, together with a smooth version (smoothed on a 5-year time scale):

Clearly there’s a lot of natural variation from year to year; the smoothed version eliminates much of the noise, making the long-term changes clearer. Another good way to reduce the noise is to take longer averages; here are 5-year averages (the last one is incomplete since we haven’t yet finished the 2005-2010 time slot), superimposed with the smoothed curve:

For land only, temperature has risen faster than the globe as a whole, due to the large thermal inertia of the oceans:

Even restricting only to land areas, the warming in the northern hemisphere has been greater still, and 2007 is by far the hottest year on record:

For the southern hemisphere the warming is quite a bit less, and the hottest year on record is not 2007 but 1998:

We can see the hemispheric contrast more clearly by plotting the two smoothed curves on the same graph:

From this we see that the slight cooling the planet’s land areas experienced from about 1940 to 1975 was pretty much restricted to the northern hemisphere; during that time, the southern hemisphere shows some wiggles but no sustained cooling. This agrees with the explanation that mid-century cooling was due to the increase of atmospheric aerosols from industrial activity. Man-made aerosols don’t stay in the atmosphere very long, so their climate impact is mainly regional (not “local,” but regional). The vast majority of mid-century industrial activity took place in the northern hemisphere, so most of the cooling effect of industrial aerosol production is confided to the northern hemisphere.

In a former post I showed that global average temperature from 1999 to the present (following the big el Nino year) was perfectly consistent with continued warming at a rate given by the trend calculated from 1975-1998 data. I also suggested in commentary that this is actually a legitimate way to inquire whether or not the trend has changed, and suggested using the 1975-2000 trend just to use round numbers. Here are the global annual averages from 1975 to the present, together with the trend line determined from monthly data from 1975 to 2000:

It’s abundantly evident that temperature since 2000 (following the regression interval) is perfectly consistent with the trend preceding it; in fact temperatures since 2000 have been a little warmer than expected from the preceding trend. We can see that even more clearly by computing the “residuals,” i.e., the difference between actual annual global average temperature and the value if it exactly followed the 1975-2000 trend:

This reveals, more clearly than any other way I’m aware of, just how wrong it is to claim that “global warming stopped in 1998.” Temperatures since then have been perfectly consistent with the preceding trend. In fact they really have been a little warmer, although the extra warmth post-2000 is not yet statistically significant.

I’m still waiting for HadCRU (the Hadley Centre for climate change research) and NCDC (the National Climate Data Center) to post their numbers for December 2007, completing the previous year’s reports. When they do, I’ll report the results.